LBNL Report Number
In an earlier analysis of the Phoenix Enhanced I/M program, we found that the effectiveness of the program decreased in the few months following final I/M testing. In this analysis we tracked individual cars over two successive I/M cycles. We find that 60% of cars that failed their initial I/M test and eventually passed a retest (presumably after repairs) passed their next biennial I/M test. However, 40% failed their next biennial test. Half of these repeat failures failed for the same combination of pollutants in each I/M cycle, suggesting that repairs were never made or were somehow deficient. More importantly, emissions of cars that passed their initial test in the first cycle increased dramatically over the two years between I/M tests; most of this increase came from older cars. This suggests that more frequent testing of older vehicles could make I/M programs more effective. The net result is that the overall emission reductions from a given fleet of cars over two biennial I/M cycles appear to have been rather small, less than 10% for each pollutant. These emission reductions are the minimum attributable to the program; reductions from what emissions would have been without the I/M program would likely have been greater. In addition, large numbers of failing cars did not appear to be completing program requirements, further limiting program effectiveness. And cars immigrating into the area from other states appeared to have slightly higher emissions than the native car fleet. I/M program managers should conduct similar analyses to determine how effective their programs are in reducing emissions, and how to address identified shortcomings.